Hello and welcome to Cogent Biosciences R&D Investor Event. My name is Valerie, and I will be your webcast engineer today. Please note that today's broadcast is being recorded. If you experience any technical issues during the broadcast, we suggest you first refresh your browser. It is now my pleasure to turn today's program over to Andrew Robbins, CEO of Cogent Biosciences. Please go ahead. Good afternoon, everyone, and thank you for joining Cogent's Research and Development Investor event. Before we dive in, I would like to remind folks that we will be looking toward the future with our comments today, and I encourage everyone to reference our recent SEC disclosures, including our 10-K for mid-March, for a much more comprehensive discussion of possible risks to our plans. Over the next hour, we plan to share with you an update on the Cogent pipeline, including details from two posters we are presenting at the American Association of Cancer Research meeting this weekend. After I make some introductory comments, Dr. Jessica Sachs will share updated preclinical data highlighting bezuclastinib's differentiation and potential to emerge as the best-in-class mutant KIT inhibitor. Then, Dr. John Robinson will provide an overview of our first novel research program, a selective potent FGFR2 inhibitor, and give you a brief look at one of our earlier stage research programs focused on ERBB2 mutations to give you a better sense for how the Cogent research team approaches drug development. Here on Slide 4, you will see a high-level description of Cogent's strategy and focus. We are an emerging leader in precision medicines for genetically defined diseases organized around our lead late-stage clinical program, bezuclastinib, which is currently enrolling patients across three clinical trials: APEX for patients with advanced systemic mastocytosis, SUMMIT for patients with non-advanced systemic mastocytosis, and PEAK for patients with imatinib-resistant gastrointestinal stromal tumors. We are looking forward to sharing initial clinical data from the APEX trial at an upcoming scientific congress this quarter. In addition, you'll see in the blue box that we have spent the past year putting together an exceptional group of scientists to create the Cogent research team, which I'll talk about more shortly. Finally, we are fortunate to be in a very solid financial position with just under $220 million on the balance sheet as of the end of 2021. Now on Slide 5, I'd like to spend some time talking about the Cogent research team, a group of experts that I believe to be as good as any small molecule chemistry-focused research team in the industry. Since April of 2021, we've managed to build a team of approximately 35 scientists with a goal of having around 50 folks in place by the end of 2022. The core of this team shares prior work experience together from our time at Array, and we continue to enhance our knowledge base by recruiting top talent from across the biopharmaceutical industry. In addition to the team based here in Boulder, Colorado, we are working closely with external partners to source dedicated, specialized talent across key functions, giving us the scale of a team of over 100 people. In parallel with building this exceptional team, we've spent the past nine months designing and building a world-class research laboratory in Boulder, which we plan to open this quarter. Moving our team into this space will dramatically accelerate our progress. As the team was coming together, they've been busy enhancing our understanding of the KIT mutant inhibitor space, including more thorough characterization of our lead program, bezuclastinib. At both EORTC in the fall of 2021 and AACR this weekend, you will see evidence of our capabilities shining light on unanswered questions about the potency, selectivity, and molecular characteristics of our program, as well as several other leading KIT inhibitors. This effort has only reinforced our conviction that bezuclastinib has the potential to be a best-in-class therapeutic option for patients with systemic mastocytosis and GIST. In addition to their work to support our lead clinical program, this team has already built its own portfolio of discovery-stage programs, including a half dozen programs focused on creating best-in-class small molecule kinase inhibitors for genetically defined oncology in rare disease patients who are in desperate need of more and better treatment options. I've been asked many times, both here at Cogent as well as when I was at Array, what is the key to such a productive and efficient small molecule research team? I believe the key to success is to assemble a group of individuals who trust and respect their peers, allow them to select targets of interest, and establish clear and realistic goals for teams to work against. To that end, as we're beginning, you will see that the majority of programs in our portfolio will fall into what we call optimized or enhanced targets. These are programs designed to develop best-in-class molecules for diseases with clearly defined biologic targets in clinical settings where the existing therapeutics are suboptimal. In some cases, these programs are designed to address patient resistance where available therapeutics cannot offer significant duration of disease control. In other settings, we are developing new molecules to improve upon drugs with significant toxicity or tolerability challenges. We believe there are many opportunities to bring new drugs forward in these classes. As we reach larger scale, we have started to add innovative targets to our portfolio, which would represent first-in-class opportunities for patient populations with poorly understood biology or for which the industry has been unable to provide therapeutic solutions. This brings me to Slide 7, which is an updated look at Cogent's pipeline. Clearly, our near-term development focus is executing on bezuclastinib's three clinical trials for SM and GIST patients. In addition to bezuclastinib, which we believe has potential to become a differentiated best-in-class KIT mutant inhibitor, we are advancing a potent selective FGFR2 inhibitor towards candidate selection later this year, which puts us on track for Cogent's first internally developed IND in the second half of 2023. We believe that there remains a significant unmet need in the market for a best-in-class FGFR selective compound which eliminates off-target toxicity while effectively covering the gatekeeper and molecular brake mutations responsible for driving several FGFR-associated diseases. In addition, you'll hear Dr. John Robinson talk briefly about our early efforts to develop an ERBB2 mutant selective for patients with specific known non-exon 20 mutations that are not well addressed with commercially or clinically available small molecule options today. With that, I'll hand the call over to Dr. Jessica Sachs, who will take us through some very exciting KIT data from our AACR bezuclastinib poster. Jessica? Thanks, Andy. I'm pleased to have the opportunity to share the new non-clinical data on bezuclastinib that Dr. Anna Guarnieri from the Cogent research team will be presenting at AACR during the afternoon poster session on Sunday 10th, April 10. Bezuclastinib is a tyrosine kinase inhibitor that is active against KIT mutations relevant to both systemic mastocytosis, or SM, and gastrointestinal stromal tumors, or GIST. KIT mutations serve as driving mutations in up to 80% of GIST patients and over 90% of systemic mastocytosis patients. The introduction of therapies targeting mutant KIT, such as imatinib, have revolutionized the management of advanced KIT, improving survival and delaying time to progression in these patients. However, although imatinib covers primary KIT mutations that arise in exons 9 and 11, these patients eventually progress due to secondary KIT mutations that occur in the ATP-binding pocket for exons 13 and 14, as well as mutations in the activation loop, exons 17 or 18. Sunitinib, which is used in second-line treatment of GIST, is active against ATP-binding pocket mutations, but not against A-loop mutations, which leaves an important mechanism for disease progression untreated. More recent advancements in the development of KIT mutant compounds have focused on systemic mastocytosis patients, as SM is a disease primarily driven by mutations in the activation loop of KIT, specifically the D816V mutation on exon 17. Unlike many point mutations for oncology indications, which represent a fraction of an overall tumor population, nearly all SM patients express the D816V mutation as a primary driver of disease. Inhibitors targeting this point mutation have shown clinical activity, but the available therapies have demonstrated tolerability challenges. To address these unmet needs in both GIST and SM, bezuclastinib has been developed as a novel type one tyrosine kinase inhibitor that is a potent and selective inhibitor of KIT A-loop mutations, including D816V. In the next few Slides, I will show you preclinical data that demonstrates why we believe this compound has the potential to be a best-in-class KIT inhibitor. First, I will show you data that demonstrates the potency of bezuclastinib. Here on Slide 10, we've used the human mast cell line HMC-1.2, which is a cell line that harbors the KIT D816V mutation, as an example of A-loop mutations, and we have compared the inhibition profile of bezuclastinib to a number of other KIT mutant inhibitors used in SM and GIST. You can see that in this model, bezuclastinib inhibits KIT activity with low nanomolar potency, 14 nanomolar, comparable to other known KIT D816V inhibitors, like avapritinib and BLU-263. That experiment demonstrates that bezuclastinib potently inhibits A-loop mutations. This data on Slide 11 demonstrates what sets bezuclastinib apart from other KIT mutant inhibitors. That is, it is exquisitely selective against other closely related kinases, such as PDGFRα, PDGFRβ, and CSF1R. These closely related kinases are sometimes called anti-targets because their inhibition represents an unintended consequence of targeting KIT and often leads to off-target toxicities. This table summarizes results from a head-to-head comparison of bezuclastinib with other KIT mutant inhibitors, and the values in the table represent the IC50 of each of these KIT inhibitors against the anti-targets shown. The color coding in the table is added to represent the activity of these compounds against these anti-targets relative to the activity of these compounds against their primary targets. For example, bezuclastinib inhibits D816V at an IC50 of 14 nanomolar. It does not inhibit PDGFR alpha at concentrations of greater than 10,000 nanomolar. It is more than 100 times more potent on D816V than on PDGFR alpha. In contrast, the avapritinib IC50 for D816V is 13 nanomolar, whereas the IC50 for PDGFR alpha and beta are 53 nanomolar and 10 nanomolar, respectively. Meaning avapritinib does not show significant selectivity for D816V over PDGFR. Since the goal of therapy is to deliver enough drug to completely inhibit D816V, the ratio of inhibition against anti-targets to inhibition against D816V would be expected to drive any toxicities observed at clinically relevant doses for KIT inhibitors. The anti-targets included here are those that are closely related to KIT, so selectivity against them is harder to achieve. Bezuclastinib is the only KIT inhibitor that has no biologic activity on these very closely related kinases. Since inhibition of these anti-targets has been linked to off-target toxicities, such as pleural effusions and edema for PDGFR and possibly cytopenias for CSF1R and FLT3, this selectivity may improve patient tolerability to treatment with bezuclastinib and appears to be a clear differentiation when compared to other D816V inhibitors. Now I want to move on to another potentially differentiating characteristic of bezuclastinib, and that is its lack of brain penetration. Slide 12 shows a non-clinical tissue distribution study to demonstrate brain exposures for three D816V inhibitors. Bezuclastinib on the left in green, avapritinib in the middle in blue, and BLU-263 on the right in red. The doses used in this experiment result in plasma exposures that are expected to be clinically relevant. After three days of dosing, which we do to better approximate brain penetration at steady state, bezuclastinib reveals a brain to plasma ratio of 0.07, thus demonstrating very low brain penetration for bezuclastinib. The observation of very low brain penetration with bezuclastinib is further supported by the absence of CNS-related activity in non-clinical safety pharmacology studies. In contrast, avapritinib has a brain to plasma ratio of two, reflecting essentially complete brain penetrance. BLU-263's brain penetrance is improved relative to avapritinib at a brain to plasma ratio of 0.18. Given the CNS-related adverse events that have been observed in clinical trials of drugs in this class, the low brain penetrance observed with bezuclastinib, coupled with its non-clinical CNS safety profile, suggests it has the potential for improved safety and tolerability at clinically effective doses. I've shown you a number of experiments now demonstrating that mutant KIT activity, which is measured by phosphorylation of the protein, is inhibited by bezuclastinib. This experiment on Slide 13 takes that data one step further to demonstrate that bezuclastinib also inhibits KIT signaling, as measured by phosphorylation of the downstream signaling protein ERK, and does so at clinically relevant doses. On the left side of this figure in green, you can see that in this in vivo model, treating HMC-1.2 tumor-bearing mice with increased concentrations of bezuclastinib resulted in decreased phosphorylation of KIT in the tumor and dose-dependent decreased phosphorylation of ERK. Complete inhibition of ERK phosphorylation is achieved at the 100 mg/kg dose, which, as you can see on the right, is a dose that achieves plasma exposures in mice comparable to the exposures achieved in our previous clinical trials of bezuclastinib and comparable to those expected at doses currently under exploration in ongoing clinical trials. To summarize, bezuclastinib inhibits KIT phosphorylation as well as KIT signaling at clinically relevant concentrations in this animal model. The experiment on Slide 14 shows that in addition to inhibiting KIT phosphorylation and signaling, bezuclastinib is also able to drive important biologic activity, including tumor regressions at clinically relevant exposures. In the left half of this figure, you can see the effect of bezuclastinib versus vehicle on tumor volume in GIST patient-derived xenograft tumor-bearing mice. At 4 mg/kg, there is statistically significant but not biologically relevant decrease in tumor growth. However, at 40 mg/kg, mice experience tumor regressions. The AUC of bezuclastinib in these mice is comparable to AUCs achieved in the clinic with 250 mg per day of bezuclastinib monotherapy. The right half of this figure shows the time course of KIT inhibition relative to dose. At 10 mg/kg of bezuclastinib, KIT inhibition is achieved at early time points, but is not sustained at 10 hours after the dose. However, at 30 mg/kg, which again is a dose that yields clinically relevant exposures, KIT inhibition is sustained for at least 10 hours in this assay. Sustained suppression of KIT activity is necessary to drive meaningful clinical benefit in patients. These data build upon the results from bezuclastinib's phase I/II clinical trial in GIST patients, underscoring the potential for therapeutic activity in patients. As I turn to Slide 15, I'll remind you that bezuclastinib is actively enrolling patients across three clinical trials. SUMMIT is a randomized, double-blind, placebo-controlled trial of bezuclastinib in patients with moderate to severe indolent systemic mastocytosis or smoldering systemic mastocytosis. PEAK is a randomized open-label phase III clinical trial of bezuclastinib in combination with sunitinib compared to sunitinib alone in patients with locally advanced, unresectable or metastatic GIST who have received prior treatment with imatinib. APEX, described here on Slide 15, is a phase II clinical trial of bezuclastinib in patients with advanced systemic mastocytosis. We remain on track to report initial clinical data on a subset of the dose optimization phase of the APEX study at a scientific conference during the first half of 2022. These data will focus on the safety and tolerability results from patients across dose levels, as well as on key biomarkers of activity, including bezuclastinib impact on serum tryptase levels in these patients. With that, I would like to turn the call to Dr. John Robinson, our Chief Scientific Officer. John? Thanks, Jessica. Today I am very excited to present the first project from our research portfolio, FGFR2. I'll start with some background information on the target and talk through how we see the unmet need and opportunity for a best-in-class FGFR2 molecule. I'll finish up with an overview of the data that we presented this weekend at AACR. The poster session is Sunday, April 10th in the afternoon from 1:30 P.M. to 5:30 P.M. The abstract number is 167. As you may know, there are four FGFR transmembrane receptors. They are inactive as monomers and function as homodimers on interaction with ligand, sorting downstream phosphorylation and signaling. There are currently three marketed pan-FGFR inhibitors, which I'll discuss in a few Slides, targeting FGFR2 fusions and FGFR3 activating mutations. The recent approvals for erdafitinib, pemigatinib, and infigratinib in patients with urothelial cancer and cholangiocarcinoma allow for use of these new drugs to improve clinical outcomes in molecularly defined patients. The challenge with this compound class by mechanism is that FGFR1 inhibition has been associated with a high level of hyperphosphatemia clinically. There are two potential opportunities here. One would be to develop an FGFR inhibitor which spares FGFR1 and possibly FGFR4 to allow dosing into a higher target engagement level in the fusion and primary mutation setting. Second opportunity will be to go after mutations that arise from on-target treatment while also sparing FGFR1. We believe we can address both of these opportunities in a single best-in-class compound. To give you a sense for how we approach drug discovery, some of the target attributes that we assess in the proof of concept stage include pre-existing therapies, emerging competition and emerging existing patient populations and subgroups. We also ask if we have expertise in a particular tumor site where we may be able to outcompete other competitors and tap into our internal expert network to find perspective that can lend insight into the unmet need as we build our target product profile. All these steps add value as we embark upon new discovery programs. Through this selection process, we have generated various datasets, such as the one you see here on Slide 18. When you perform an assessment of FGFR2 and FGFR3 alterations together across all cancer types, you have about 4% overall alteration rate. Of that 4%, roughly 80% are FGFR mutations with 8% amplifications, some fusions, and there are 3% that have multiple mutations. Today I will focus on the top right-hand box where we talk about FGFR2 altered cancers. As you can see across tumor type, there are a large number of mutated patients at initial diagnosis. The question then becomes, are there other therapeutic options which are competitive in this space, or is this mechanism by far the best or only therapeutic option that exists? In this case, FGFR2/3 inhibitors would be the preferred option. What I'd like to point out here is that there are a lot of potential indications for both FGFR2 and/or FGFR3 inhibitors. As an example of clinical opportunity and experience, and to better understand how we might differentiate as a best-in-class compound, I want to talk a little bit about cholangiocarcinoma. On the previous Slide, I showed that there are a variety of mutations and fusions across tumor types for FGFR2. For cholangiocarcinoma, the initial opportunity for first-in-class drugs, which are in fact pan inhibitors, is to treat extracellular domain mutations, which are shown in green here at the left side of this figure. If you look at mutational status at time of diagnosis, there are a large portion that have extracellular domain mutations. Though the relative percentage is low, there are also a variety of other kinase domain mutations that exist at diagnosis, including the molecular break N549 shown on the right. Clinical efficacy is predominantly driven by the ability to suppress FGFR2 extracellular domain mutation-based signaling. The same story is also true for FGFR3 in urothelial carcinoma. One of the most predominant arising mutations from treatment with pemigatinib, erdafitinib, and infigratinib is the gatekeeper mutation V564, for which there is no current therapy. As you can see from the box stat graph on the right, in the second-line setting, many patients have multiple mutations, and the largest opportunities are emergent molecular break and gatekeeper mutation. As you can see from the three marketed pan-FGFR inhibitors, there is a clear and clinically approved treatment benefit for FGFR wild-type extracellular domain mutated patients. One of the largest challenges associated with the clinical progression of these compounds has been hyperphosphatemia. What has been widely accepted in both the research as well as the clinical community is that FGFR1 is the primary driver of hyperphosphatemia and the largest single contributor to dose schedule change and interruption. None of the three marketed pan-FGFR inhibitors have shown appreciable activity against kinase domain activating mutations for FGFR2 or FGFR3, nor resistance mutations for FGFR2 or FGFR3. The simple hypothesis for a best-in-class first to target compound would be a drug that spares FGFR1 to accommodate dosing to engage for kinase domain activating and resistance mutations in both patient populations, for which there is no current mutation-targeted therapeutic option. In addition, dosing holds and dose reductions are extremely common with the first-generation pan-FGFR inhibitors in order to manage the FGFR1-mediated hyperphosphatemia, and these dose interruptions and reductions may lead to incomplete target engagement. This leads us to the position that there is additional efficacy and potential clinical durability that remains untapped due to non-optimal long-term engagement in the first-line setting. Let's take a minute to familiarize ourselves with the ATP binding site of wild-type FGFR2, in this case with imatinib bound, as an example of how we think we can navigate the binding site to make compounds that allow us to target both FGFR2 gatekeeper and resistance mutations. As you can see, the normal gatekeeper, valine 564, shown in yellow here, does not impinge into the pocket that the tetra-substituted phenyl ring occupies, which is the hydrophobic pocket or space in the far back in this orientation. As the gatekeeper mutations occur, commonly substituting phenylalanine or isoleucine for valine, which are much larger amino acids on a relative scale, the steric space toward the back of the pocket becomes compressed. Development compounds that have large substituents in the area circled in red of the infigratinib structure are no longer able to bind the active site. The molecular brake mutation, N549H, occurs outside the ATP active site. Since the molecular brake mutation does not directly affect the ATP cleft sterically, many of the first generation compounds are somewhat potent against it, but are limited in clinical utility by target engagement of FGFR1. As lead generation projects begin at Cogent, there are multiple routes for us to generate novel lead series. One path is through structure-based design, beginning with modifications of a hit from a screening collection. A second is de novo structural design based on our understanding of the binding site and SAR of available compounds. This project was de novo design-based, and we were able to quickly find novel chemical space and develop structure-activity relationships in multiple series. The initial project proof of concept goal was to elucidate a clear lead series, or ideally two. This was accomplished by examining multiple core changes and novel design elements that led us to progressible chemical space. At the onset, we believed that we could reach a robust lead series if we could develop SAR of many analogs within a series that exhibited thirty-fold or greater selectivity versus FGFR1, with a variety of different chemical moieties employed as SAR points. As you can see from this Slide, we've been able to synthesize a broad potency range of compounds with accompanying selectivity for FGFR2 over FGFR1. I would say the majority of our compounds are at least 30-fold selective for FGFR2 over FGFR1, with a clear path to compounds which can significantly outperform in this selectivity. To date, we have generated multiple lead series that are both potent and selective for FGFR2 versus FGFR1, and I will describe one series exemplified by CGT-0292 over the next few slides. As you saw in the previous slide, we didn't necessarily select the most potent or most highly selective compound as an exemplar to use in profiling our cell-based assays, as well as to develop our models. One of the reasons that we chose this compound, CGT-0292, is because it has physicochemical and drug-like properties that allow us to provide validation for the lead series in vivo and to provide robust avenues for further optimization. As you can see from the summary slide of data in this table, CGT-0292 is selective for FGFR2 over FGFR1. It exhibits BCS class 1 compound properties with a solubility of greater than 300 nanograms per mL across pH range, high permeability as measured by MDCK-1 assays, with PGP efflux ratios that should favor CNS exclusion, protein binding in the 30%+ range across species, in addition to a relatively low level of predicted metabolism across species, which is predicted by both microsomes and hepatocyte. We also address reversibility in the enzyme assay with preincubation to show definitively that CGT-0292 is a reversible inhibitor. This avoids potential for time-dependent inhibition of FGFR1 or any other off-target kinase or receptor that we may have slight activity against down the road. It also removes the potential for clinical variability through differentiated glutathione metabolism amongst patients. Based on the overall in vitro profile and performance to be discussed on subsequent Slides, we believe that CGT-0292 represents a robust scaffold and lead series that can be further optimized. One of the biggest goals of the project to date has been to establish a firm selectivity window of FGFR2 over FGFR1. As you can see here, when we compare CGT-0292 to infigratinib and the rest of the field of pan-FGFR inhibitors, we have drastically increased the window for FGFR2 target engagement versus FGFR1 wild type. CGT-0292 is also potent against the FGFR2 gatekeeper and molecular brake mutations, as well as FGFR3. I would like to reiterate here that this level of selectivity over FGFR1 wild type could have long-term potential advantage, not only in the resistance setting, but also in terms of ability to potentially provide a more robust molecular response for FGFR2 and FGFR3 fusion patients. After we've established our in vitro data set with respect to in vivo metabolism, one of the first things we like to do is assess oral performance in at least one rodent species. Ideally, we will look at both mouse and rat. As you can see from the PK graph on the left, CGT-0292 exhibits increased exposure with dose over the 10-100 mg/kg range, as well as medium observed clearance. The actual measured clearance value here is around 36 mL/min/kg, which is 41% extraction ratio. This correlates reasonably well with the microsomal and hepatocyte predicted performance. For reference, the 10 mg/kg dose is 88% orally bioavailable. On the right side is this new SNU-16 xenograft model. This is an FGFR-directed amplification model and cell line, where we assess the ability of CGT-0292 and infigratinib to suppress phospho-FGFR2 signaling. As you can see, at 30 mg/kg, CGT-0292 completely suppresses phospho-FGFR2 signaling. Infigratinib performs equally well at 100 mg/kg. It's important to note that we chose to dose infigratinib to achieve levels above the highest achievable clinical Cmax, as we needed to validate the dynamic range of the PD model by completely suppressing the phospho-FGFR signal. Here you can see equal performance of CGT-0292 and infigratinib both at relatively high dosing. As a second data point to address in vitro to in vivo correlation with respect to metabolism, we dose escalated CGT-0292 in rats. Between 10 and 100 mg/kg, CGT-0292 exhibits dose proportional exposure. Again, we observe well-behaved overall performance, as we see roughly 86% oral bioavailability at 10 mg/kg. The observed clearance here is 19.4 ml/min/kg, which is around 28% extraction ratio, which is predicted by low but in vitro clearance in both hepatocytes and microsomes. This compound performs well in both mouse and rat and has reasonably well-aligned in vitro to in vivo predicted clearance. As a test compound for this series, this is an important aspect for us to assess, as well-behaved IVIVC increases our confidence within the series toward further optimization and eventual delivery of a drug candidate. At the right is a rat model of hyperphosphatemia. As you can see, we compare doses of CGT-0292 between 50 and 20 mg per kg to that of infigratinib at 60. Again, we dosed infigratinib to a level where we believe we should be able to observe hyperphosphatemia based on clinical data. This was done both to validate the model and, more importantly, to be used toward future optimizations and predictions of therapeutic window. This is dose and Cmax time match for rodent exposure and plasma protein binding versus FGFR1 free cellular IC50. The take-home message here is that extended exposure around or above three fraction adjusted IC50 for FGFR1 will lead to a rodent response with respect to hyperphosphatemia, and we believe this is a well-behaved rodent response correlative versus clinical hyperphosphatemia, which has been used and described by many companies. To summarize, these data show that we can develop a compound with highly differentiated selectivity and that we are on a clear path to generating additional compounds with enhanced selectivity to maximize therapeutic activity against FGFR2 without driving FGFR1-mediated hyperphosphatemia. Over the past few slides, I've shown you the importance of navigating FGFR1-mediated hyperphosphatemia clinically, and that related toxicology associated with pan inhibitors is largely related to FGFR1. I've described CGT-0292 as an exemplar of our research efforts, demonstrating that it can provide a 30-fold or greater selectivity window for FGFR2 over FGFR1, that it is potent against the panel of FGFR gatekeeper and molecular brake mutations, and it has highly favorable drug-like properties and in vivo performance. We've established a robust FGFR2 PK/PD model, as well as a model to assess therapeutic window versus hyperphosphatemia. We are currently working on further optimization of CGT-0292 and the lead series to expand our therapeutic window versus hyperphosphatemia with the goal of delivering a clinical candidate into IND-enabling studies toward the end of 2022. Now let's switch gears to our second target in our early research portfolio, ERBB2. As many of you may know, ERBB2 is a receptor tyrosine kinase that belongs to the ERBB2 EGFR four family. It's also commonly referred to as HER one through four. This is the epidermal growth factor receptor family receptor tyrosine kinases. HER2 is aberrantly expressed and activated in many differentiated cancer types. The most common occurrences are in breast, where the Pertuzumab/Trastuzumab antibody combo has now become standard of care. In addition to the antibodies which block receptor dimerization or promote receptor internalization or degradation, there are also ADCs and small molecule inhibitors that are available treatment options for HER2-driven malignancies. Currently, the only FDA-approved small molecule inhibitor that is selective for ERBB2 in this space is Tucatinib. The historical liability of small molecules in this space is inhibition of or lack of selectivity versus EGFR. There are at least two distinct opportunities linked to ERBB2 that have yet to be well explored clinically. The first is exon 20 insertion mutations. This class of mutation is very common in metastatic breast cancer and lung cancer with a high level of CNS involvement, and unsurprisingly, the competitive space is quite crowded. The second class of mutations, those that are non-exon 20 mutations, occur in a variety of cancer types outside of breast and lung. These mutations include the S310F, R678Q, L755S, V842I, and D769Y. As you can see from this table, there are quite a large number of non-exon 20 mutations that emerge per year in bladder, urothelial, and stomach adenocarcinomas as examples. Currently, there are no ERBB2 mutant-selective therapeutic options available to treat these patients, and unlike exon 20 approaches, there are relatively few competitors pursuing molecules for these patients. An additional point here is that amplification and mutations are in a mutually exclusive fashion, representing independent drivers of cancer progression. Cogent is currently working to describe potent selective small molecules to target these non-exon 20 ERBB2 mutations, which remain a significant unmet need for these patients. To further dig into this potential opportunity, we've looked across both TCGA and GENIE databases to assess the relative mutational frequency of each of the non-exon 20 mutations. Not surprisingly, there's a low relative frequency of gatekeeper mutations, as to date, very few of these patients are treated with small molecule ERBB2 inhibitors. There are a large population of S310F and S310Y mutations observed, which drives ligand-independent constitutive dimerization. To best understand the opportunity, we then profiled a variety of standard EGFR inhibitors as well as pan EGFR ERBB2 inhibitors and novel selective series across a panel of mutant cell lines related to the tumor types described on the previous Slide. As you can see, many of these compounds appear to be robustly engaging the non-exon 20 ERBB2 mutants across the panel, highlighted in green in vitro. Highlighted in red, column one, you'll note that many compounds hit EGFR wild type at relatively similar potency levels versus the ERBB2 panel. In this context, coverage of wild-type EGFR limits these compounds clinically with respect to their utility versus the desired ERBB2 mutations. While Tucatinib and the related compound BI-1622 appear to have a reasonable selectivity window over EGFR, it's important to note that this is only in vitro in a cellular context, which is a critical point we will discuss in a moment. We selected Mobocertinib and Tucatinib for further assessment of clinical target engagement based on available data. In this table, we assess Tucatinib and Mobocertinib based on their clinical Cmax target engagements versus the mutant of interest S310F, S310Y, L755S, and V842I. Tucatinib has a well-established clinical Cmax concentration of around 581 nanograms per ml, which correlates to 1,210 nanomolar clinical Cmax plasma concentration. Since Tucatinib is 2.9% free, that leads us to a free fraction-adjusted IC50 for L755S of around 4 micromolar. Since we believe these mutations are driven clinically, it's likely that we'll need IC90 coverage for robust clinical response. As you can see from the column at the far right, neither Tucatinib nor Mobocertinib can cover any of these mutations appreciably above 3 cellular IC50. My perspective is that this is a clear opportunity to provide a novel therapeutic option to treat patients with these mutations, as none of the current approved therapies can nor have been used to treat successfully. On the previous slide, we highlighted BI- 1622, which is highly structurally homologous to Tucatinib. Since we believe that some of the clinical liabilities of Tucatinib with respect to dose ascendability are closely correlated with physicochemical properties of this chemotype, there is a clear opportunity to both differentiate and outperform in this space. That leads us to our target product profile shown here on Slide 35. The goal is to develop an ERBB2 mutant-selective compound that is at least 20-fold selective for EGFR wild type. We are designing compounds to be well-behaved with respect to physical properties, have dose-ascendable performance, have no liability with respect to co-administered medications. As noted at the bottom of the slide, CNS penetration would be an upside, but it's not a primary goal of developing a drug for treatment of systemic disease driven by these ERBB2 mutations. What do we have that's different or unique from all of the other approaches that have been played out historically? As this is a de novo design project, we realized early on that a huge advantage here would be the ability to solve crystal structures with any of the mutants of ERBB2, as it has been notoriously challenging to obtain structural data due to protein stability and isolation issues. This is reflected in the public domain data, as there are very few ERBB2 structures versus a large number for EGFR. What you can see here on Slide 36 in the left panel is that we've been able to to stably express, modify and isolate the V842I mutant protein bound to a covalent inhibitor. We were then able to prove single site modification for the V842I mutant protein, highlighted in the mass spec assessment in the middle panel. In the next frame, we've been able to grow crystals of the ERBB2 842 mutant with the covalent inhibitor, and we're currently in the process of solving the crystal structure for this series of compounds. We believe that this is a huge advantage for Cogent as we move the project forward, as this is the first example of ERBB2 mutant structural information that we know of. Currently, our ERBB2 selective project is in lead generation. We've identified multiple novel series of compounds that have the requisite selectivity over EGFR wild type and has some of the driver mutants of interest. Our focus is now to drive towards selection of two lead series focused on maintaining selectivity over EGFR wild type while increasing potency against ERBB2 mutants. As you can see from the table at the bottom, we have two distinct series that are selective versus EGFR. The series are structurally diverse, including both covalent and reversible inhibitor classes. We expect that in the next quarter, we will enter into lead optimization for multiple series of compounds. With that, I'll turn the call back to Andy. Thanks, John and Jessica, for walking us through the excellent progress the team has made in a short period of time. At this point, I'd just like to summarize where Cogent is at as a company. We are currently enrolling three clinical trials for our lead program, bezuclastinib, which we believe has the potential to become the best-in-class KIT mutant inhibitor for patients with systemic mastocytosis in imatinib-resistant GIST. This is based on results from bezuclastinib non-clinical studies, which showed to be a highly potent, highly selective, non-brain penetrant KIT mutant inhibitor, supported by its activity and tolerability profile from a 50-patient phase I/II clinical trial. We are looking forward to sharing initial clinical data from the APEX study in advanced systemic mastocytosis patients this quarter, focused on its clinical activity and tolerability profile in patients. Looking toward the future, we are excited to share that Cogent is building out a robust pipeline of small molecule inhibitor programs targeting genetically defined diseases. Starting with our FGFR program, our plan is to create a molecule that will both spare FGFR1 inhibition and related toxicity, as well as potently cover the relevant molecular breakpoint and gatekeeper mutations associated with this target. Based on our progress, as exemplified by CGT-292, we feel confident that this program will deliver an IND during the second half of 2023. We are also excited today to share an early look at our non-exon 20 ERBB2 mutant program, which has the potential to provide a potent and selective therapeutic option for these patients with a significant unmet medical need. The team has made great progress, including recently solving a key protein crystal structure, which will accelerate the timelines for our drug discovery program. As other programs progress from early discovery into lead generation, we will identify opportunities to share with investors more details on those exciting programs, including some true first-to-market opportunities. Before turning the call over for questions, I'd like to sincerely thank the exceptional Cogent team for their incredible performance to date, and I think I speak for the entire patient, investigator, and investment community when I tell them that we are all excited to see what they can do next. For that, I'll open the call for questions. Operator, you may now open the line for questions. Thank you. Ladies and gentlemen, if you'd like to ask a question, please press star, then one on your touch-tone telephone. Again, if you'd like to ask a question, please press star, then one. Our first question comes from Eun Yang of Jefferies. Your line is open. Thank you. T he next data catalyst is the phase II APEX data in advanced SM in Q2. Now you are enrolling patients in phase II SUMMIT study in non-advanced SM patients. When do you think we would see some preliminary data in non-SM patients? Second question is on this ERBB2 mutant selective program. It's earlier than FGFR2. When do you think you would be in a position to move into IND filing studies? Thank you. Great. Thanks, Eun, for the questions. Appreciate it. F rom your first question about the timing on SUMMIT, which is, as you said, our phase II study for patients with non-advanced systemic mastocytosis. The design of that study, I'll just remind folks, is split into essentially two parts. The first part is a double-blind randomized trial that is going to enroll non-advanced patients across either three different doses of bezuclastinib or to placebo. The purpose of that, of course, is to try to pick the appropriate dose to take forward into part two of the study, which would be designed to support registration if successful. It's also designed to help validate a patient-reported outcome instrument that we would be using as the assessment of the primary endpoint during part two. It is very important for us to keep part one as a blinded study for the purposes of that second part of part one that I just went through. We do need to enroll 48 patients in part one of SUMMIT, and then we need to follow them up for a significant duration to assess both the activity and safety and tolerability of the drug, but also, again, validate that instrument. While we haven't put a stake in the ground for exactly when that's going to occur, I think we have said that it's unlikely to occur in 2022, so it's probably more of a 2023 event. As to your second question on the timeline for the ERBB2 program, as you alluded to, it is earlier stage in the discovery process than the FGF program that John walked through. It's probably a little early for us to, again, peg a specific time, but based on the scale, the size of the team that we're building, the expectations and excitement that we have around the Cogent research team, I do think we want to head towards a goal of trying to deliver a novel IND program to the clinic every year. We're probably not in a position to say a specific year, but we think FGFR2 will come first in the second half of 2023, and then we'll see how the ERBB2 mutant program progresses over the next 6-12 months before putting a firm stake in the ground. Thank you. Thank you. Our next question comes from Chris Raymond of Piper Sandler. Your line is open. Thanks for doing this comprehensive overview. I appreciate all the info. I've got two questions. First, maybe for John on the FGFR2 program. G uys, I'm hearing you on the deleterious effects of hitting FGFR1, but curious if you had a thought on this. T here are other competitors in the space that have argued you need to hit additional FGFRs, including FGFR1, to block compensatory mutations that arise through FGFR oncogenic signaling. Just maybe your thoughts on this. Do you see this as an issue or maybe any sense as to that thesis from your perspective? On the APEX data that's anticipated later this year, maybe just remind us. I know this is a dose ranging study, but give a sense of what you're hoping for in terms of tryptase reduction, in particularly in the context of what we saw with AYVAKIT. T hey saw relatively fast and pretty sizable tryptase reduction even at lower doses. Should we expect something the same, or some other pattern? Thanks. Great. Thanks, Chris. John, I'll take the second question first, and then maybe I'll turn it over to you for your thoughts on that first question. With regard to APEX and the clinical readout, it's unfortunately not as easy for us to say what you need to look for is a target tryptase reduction of below or above X. Part of that is due to the fact that we're exploring in part one of APEX several different doses, ranging from between 100 to 400 mg of daily bezuclastinib. We're also allowing for patients who both have and have not seen prior avapritinib to enroll in our study. You might expect that a patient receiving a 100 mg daily dose of bezuclastinib that had seen prior avapritinib might expect a different outcome on clinical activity versus a patient that was receiving the highest dose, daily dose of bezuclastinib and had no prior exposure to another potent KIT inhibitor. What we've said is that when we get to the clinical readout for whatever number of patients we have in the study at that point, we will describe on an individual patient basis a time series so that it'll be obvious to folks how bezuclastinib is impacting serum tryptase. As you said, it's not a metric that you need to wait 6 or 12 or 18 months to get a good read on. That's why it is useful in this disease as a surrogate. I think it'll be, as I've said before, probably interpretable data that you will be able to tell at this initial data readout is bezuclastinib impacting serum tryptase to a material fashion. John, do you want to give any thoughts on the question about the benefits of involving FGFR1 on? I'd say it's a great question, and certainly we've thought about this a bit, and we're apprised of the positioning of other competitors in this, I'm going to say, the second-line space, right? I think for us, the big focus right now is to make sure that we maintain window on FGFR1, so we can navigate hyperphosphatemia clinically. We have, as you saw from, I don't remember the slide number, I think the slide where I described our initial SAR efforts. We have compounds that cover a variety of different FGFRs as well as mutants. As part of the tolerability and efficacy assessment as we move compounds forward, we'll continue to think about that. That's a good point. Right now, the main drive is to have a compound that can provide a deep molecular response in that resistant setting, as well as maybe a better response in first line and steer clear of FGFR1. Great. Thanks so much. Andy, do you have any other comments on there? Thank you. Our next question comes from Sam Slutsky of LifeSci Capital. Your line is open. Hey, thanks for the questions, everyone. A couple for me. Fi rst, can you just walk us through the process you took to choosing FGFR2 as the first target to pursue, mutant ERBB2 as the second, and then going forward, how should we think about what potential future targets might look like? Hey, Sam. Sure. I can give it a shot, and then John can tell me that I didn't listen very well. When we try to select targets, we look for opportunities where we believe we can develop molecules that have, I guess, physicochemical properties to be drugs and to bring both clinical activity to the patient population as defined by some genetic mutation. C learly, those are both in scope. We also try to look for opportunities where the competitive space doesn't include two dozen other companies, including six or eight big pharmaceutical companies who are further down the road, so that we have to play catch up with so many competitors to create these compounds. We think, again, FGFR is probably a great example of this. While we're certainly aware that there are other biotech companies who are pursuing what I'll call loosely FGFR selective programs. We sat down and we thought through each of those efforts that we're aware of and how we could potentially differentiate, and we think that we have a specific story about why, for instance, a reversible FGFR2 selective that completely spares FGFR1, what might become best- in- class in this field. From an ERBB2 perspective, it's not lost on us that there are many companies who believe that creating a new small molecule that can cover exon 20 mutations is exciting, and so we wanted to acknowledge that. In the family of ERBB2 mutations, we also were aware, working with our external scientific advisory board, that there were many well-characterized, known point mutations that drive cancer types that are underserved with small molecule ERBB2 inhibitors, either because they have activity against EGFR wild-type or because you can't provide a therapeutic index that gives you clinical exposure sufficient to cover the target. Again, it's a great example of limited competition, well understood biology in a space that we think from the Cogent research team's core competency, we can design chemical matter that can address those specifically. Maybe I'll pause and see, John, if you want to add anything about those two. No, I agree with you. I would say, when we started the company, we looked at more than two. W e didn't start with just these two. We looked at six or seven different opportunities. Some of those are listed as targets three and four at the bottom, and then there were probably five other ones that we down selected. W e looked across different opportunities, and as Andy pointed out, for the reasons and attributes that he's already discussed those two were selected. Sam, just to finish off the end of your question, and sort of building on John's comments, we're not going to stop with two programs. Obviously, we're working on several other small molecule kinase targets in Hit ID, which I think we call the first phase of our discovery program. T he way that I characterized them before between enhance, optimize, and innovate, at least a few of those targets are, let's call them, more classic undruggables. They'll probably take a little bit longer because those programs often take more time to really suss out and figure out exactly how you want to approach them. But they admittedly will have, I would say, significantly larger commercial potential down the road. We also felt it was important, as a biotech company of our size in our maturity, we've only been around for 18 months, to go out and demonstrate our capabilities with a company proof of concept that we can build molecules that have the potential to become best in class. I truly believe that our FGFR and our ERBB2 programs will be able to deliver that over time. W e now have to go out and prove it. Got it. Okay. I guess for FGFR2, can you just remind me what the on-target toxicities are for a specific FGFR2 inhibitor? I have a follow-up or two. Yes. I think the one that folks talk about frequently is ocular toxicity, so retinal pigment epithelial detachment, which is associated with inhibiting FGFR2. I would say, from the clinical experience put up by our colleagues at Relay with RLY-4008, they demonstrated, I would say, high rates of stomatitis and hand-foot that maybe were unexpected. I think that there's an ongoing discussion around whether those are FGFR2 on target toxicities or whether that has something to do with time-based tolerability challenges of covalent approaches and irreversibility, potentially inhibiting other kinases that might look selective in an enzymatic or cellular assay. When you leave drug on board semi-indefinitely, you may accumulate toxicities outside of intended consequences. I think there's an open question. We of course are going to be mindful of things like the ocular toxicity associated with FGFR2, but we'll continue to monitor the space and see how that unfolds. Got it. I guess in the sense that you're creating a reversible FGFR2 inhibitor, to your point, Relay Therapeutics, and can you just walk through thoughts around increasing therapeutic index and potential advantages of being reversible in this case? Yes, I think that sounds like a great question for John. Do you want to try to give a sense for in the world of FGFR inhibitors, why you believe reversibility could turn out to be an advantage longer term? I think the one big question about potential toxicology associated with off target is always something to think about. I think time-dependent inhibition of FGFR1 and/or other closely related kinases or receptors they may hit could potentially be a liability. You know what, we have a little bit of experience from a previous life with other covalent compounds, and I think that differentiated glutathione-based metabolism leads to some reasonably large CV amongst patients as you go forward. A nother thing that we spent a decent amount of time talking about internally is proteome selectivity, right? W hen you have that covalent modifier around at a high local concentration, in the GI tract, you can have off-target effects from that because since it's covalently modifying, once the protein is, I'm going to say, "engaged or off", that can lead to this long-term tolerability or tox consequence. There's a couple of advantages, in that way, of thinking. It's all about time-dependent inhibition of the target or additivity of pharmacodynamic effect, however you want to think about it. Okay. Just the last question from me, just for the ERBB2 mutants that you're targeting, what's currently used in these patients today? What do outcomes look like for these groups? Yes. I t's a good question, and we've asked lots of folks who treat these patients that question. It's probably a different question whether you're talking about a gastric or a uterine patient or bladder cancer patient who gets their treatment at a NCCN center where NGS is frequent and ubiquitous versus at a community cancer center. In general, I would say the vast majority of community patients that have these diseases are probably treated with standards of care, chemo plus minus IO as they move through lines of therapy. I would say in certain pockets, again, at major NCCN centers, I believe that these patients, they're trying things like tucatinib and lapatinib and neratinib when they see ERBB2 mutations pop up. I think that there's a general understanding that those drugs don't seem to be active in these patients, so it's never going to get a real hold in place. Part of this project will be to help educate treaters of lower abdomen organ primary cancers that it's important to do sequencing. When these mutations pop up, if we can build a drug and we can demonstrate it clinically, then we will be part of creating a new market for those patients. Cool. Thanks for the questions. Sure. Thank you. Thank you. Our next question comes from David Nierengarten of Wedbush Securities. Your line is open. Taking a question maybe a bit more on the rise of resistance in FGFR2 and related cancers. I t appears to be, you could have an advantage with a reversible inhibitor, but then the counterargument would be that the irreversible inhibitors that we see on the market have been disappointing in terms of duration of response. Would a reversible inhibitor allow resistance to rise even faster to your candidate? D o you think that that's a n important question, or do you think that maybe the other drugs with their FGFR1 activity are just suboptimally dosed? H ow do you see that interplay going forward and with the dialing up or dialing down activity, and such on the drug candidate? Yes. I think that's the $64,000 question is if you have a drug that is bringing along FGFR1 and you're willing to live with hyperphosphatemia, well then what is the dose limitation on being able to achieve IC90, and for how long can you do that before you have to back off because of some of the consequences of having hyperphosphatemia in a person? T Hen on the other side, I think your question is if you have a reversible inhibitor, there's lots of reversible inhibitors that are known out there, and many of them confer resistance in patients. I would just say there's a spectrum, right? I'm sure we can pick our favorite reversible kinase inhibitors that have quick resistance, and we can pick our favorite ones that have a significant duration of therapy before resistance arises. A t the stage we're at, it's hard to predict exactly where on that spectrum our hypothetical drug would fall. We do believe that some of the experiments are either being tested currently clinically, for instance, Relay's irreversible inhibitor. We think that other versions of this question will be tested at some point via Kineta and Tyra, which have, let's call them, maybe slightly more selective, but still FGFR1-involved pan-FGFR inhibitors. We do like the space, not only just intellectually but from talking to lots of treaters of patients with FGFR mutations of creating a small molecule that does have significant step-up in selectivity between two and one, but also has that benefit of being reversible. I think we can carve out that space, at least, in our opinion. Again, we understand the need to move with some urgency to get a drug candidate and then move into the clinic. We have a hypothesis that that could become best in class. Again, we can't promise you that, at this point. Sorry, just a quick follow-up. Is it known if FGFR2 naturally generates mutations faster? I s it just a more mutable kinase than maybe some of the others that c an be inhibited for a longer time or have b etter duration of therapies and different kinases and different cancers? I think it's probably hard to answer that question completely because there probably are not... The pan FGFRs that are commercially available. I think they implicate FGFR2, you certainly based on some of the response rates that you've seen in clinical trials. It's tough to say that they've achieved an optimal therapeutic index for covering FGFR2 because some of the tolerability challenges of bringing along some of the other isoforms potently. I think it's going to take somebody creating a selective FGFR2 to directly answer that question with confidence. I think the jury remains out. Okay. Thank you. Sure. Thank you. Our last question will come from Matthew Kaplan of Ladenburg Thalmann. Your line is open. Hi. Thanks for taking the question and, thanks for the presentation today. I guess a couple things, first on APEX. With respect to the interim data, I expected you gave some good detail in terms of what to look for on the serum tryptase, but can you give us a sense in terms of the duration of treatment and, what type of sense we'll be able to see in terms of tolerability profile on the interim look? Yes. Until we actually do the presentation, it's tough to speak to duration of therapy. But again, I'll just reiterate, we believe that when we present these initial clinical results, they will be interpretable, so w e believe that they'll bring some good data to the forefront on how bezuclastinib's behaving and performing in these patients. From a tolerability perspective, it's certainly a cornerstone of differentiation between bezuclastinib and avapritinib, and we hope that long-term improved tolerability can translate into improved therapeutic index and hopefully improved clinical activity and duration of therapy. But that's going to take a significant amount of time because, as I think we all saw in PATHFINDER, avapritinib is an active drug. I don't think there's any doubt about it. It's just probably hard to tolerate for these patients. From a safety and tolerability perspective, there are, I think, high-incidence adverse events that I would direct folks to look for right away, like the PDGFR-linked edema that avapritinib, I think, falls at around north of 70% incidence across their large trials. Because bezuclastinib does not seem to implicate PDGFR, one would expect, if that is demonstrated clinically, you would see a significantly lower rate of edema, and you would be able to tell that even in a relatively small number of patients. On things like cytopenias, on cognitive impairment, and all the way to things like intracranial bleeding, there's a spectrum on frequency, depending on which trial you're trying to do a comparison against. Obviously, if we have a few handfuls of patients, it's going to be hard for us to definitively state that because we saw no intracranial bleeds, you can put that to bed. I would hope that we don't show any. On things like cognitive impairment and the rate of cytopenias, I think that there is an opportunity to start to tell a story about differentiation as well. The other thing I'll add is because it's an open-label study, and we'll continue to accrue patients, we're not necessarily slowing patient enrollment for a clinical data readout. We'll have the opportunity on a relatively regular basis to share updates from the APEX trial and continue to build the body of evidence that this is a well-tolerated and potential best-in-class KIT inhibitor. Okay. That's very helpful. Thank you. Secondly, on the PEAK study, the GIST phase III study, can you give us an update on how that study's progressing? Yes. The PEAK study, I think what we say across all these is they're actively enrolling, and we don't necessarily give updates about there being X patients on every Tuesday. But what I can tell you is that the design of the PEAK trial starts with a lead-in that is designed for us to prove to ourselves and patients and investigators, but also importantly the agency, that an updated formulation of bezuclastinib that we spent most of 2021 creating has predictable pharmacokinetics and exposure relative to the original formulation that we used in the phase I/II study. We are going to have to enroll, or we want to enroll, 12-18 patients in the lead-in to the PEAK study in order to ensure that we're giving similar exposures to what we saw as clinically relevant doses in the phase I/II trial. Once that's complete, the way that we've organized the study, we can move immediately into the randomized phase. What's called the first portion of 2022 is going to be spent doing that lead-in exploration to make sure that we understand exactly the behavior between the original formulation and the updated formulation. Then at some point in mid-2022, I think we've said before, we plan to move into the formal randomized portion of the global phase III trial. When we get to that point, we'll be in a much better position to tell folks about specific dose and number of pills. What we said before is we're quite confident based on the updated formulations, performance in our internal models, that we will significantly reduce the number of pills required to get to the exposures we saw in the phase I/II trial. Okay, great. Well, thanks for taking the questions. Thank you. Thanks to all the folks who joined the call and asked the questions, and thanks to John and Jessica for the presentations. At this point, I think I'll turn it back over to the operator. Thank you. Thanks to all the participants for joining the call today. This concludes our webcast. You may now disconnect. Have a great day.